emub
asked on
Building a Tree from DataTable results with the node structure formatted as \Node1\Node2\
Hi,
I've been banging my head against a wall for the past few days with the problem of turning DataTable results into a Tree Structure to populate a TreeView control. The tree path is stored in a field called Path.
[Path]
A\1\A\1
A\1\A\2
A\1\B\1
A\1\B\2
A\1\C\1
A\2\A\1
A\2\A\2
A\2\B\1
B\1\A\1
C\1\A\1
Normally I do a recursive query using parent ID's but this system is different as the Root Parents Childs are all stored in the string. I've been doing a For each Row in DataSet and breaking down each Path into a string array using NodeList = Path.Split(CChar("\")) but its the building of the nodes that I just cant get my head around.
Idealy I dont want to requery the database again, the results in the datatable from a stored procedure with a fair few params. So I'd have to work with data just in the datatable.
Regards
Daniel
I've been banging my head against a wall for the past few days with the problem of turning DataTable results into a Tree Structure to populate a TreeView control. The tree path is stored in a field called Path.
[Path]
A\1\A\1
A\1\A\2
A\1\B\1
A\1\B\2
A\1\C\1
A\2\A\1
A\2\A\2
A\2\B\1
B\1\A\1
C\1\A\1
Normally I do a recursive query using parent ID's but this system is different as the Root Parents Childs are all stored in the string. I've been doing a For each Row in DataSet and breaking down each Path into a string array using NodeList = Path.Split(CChar("\")) but its the building of the nodes that I just cant get my head around.
Idealy I dont want to requery the database again, the results in the datatable from a stored procedure with a fair few params. So I'd have to work with data just in the datatable.
Regards
Daniel
ASKER
Though I think the toughest problem isnt the recursive part its working with the path.
Here is the code I've done so far. it almost works but I cant work out why it gets stuck in a loop.
Tracing through it would make the following nodes
D
D\1
D\1\A
D\1\A\1
D\1\A\2
D\1\A\3
D\1\A\4
then it would start again back at D\1\A point and loop forever.
Here is the code I've done so far. it almost works but I cant work out why it gets stuck in a loop.
Tracing through it would make the following nodes
D
D\1
D\1\A
D\1\A\1
D\1\A\2
D\1\A\3
D\1\A\4
then it would start again back at D\1\A point and loop forever.
LocationTreeView.Nodes.Clear()
For Each root As String In GetNodeRoots(locationList)
LocationTreeView.Nodes.Add(BuildNode(root, locationList))
Next
Public Function GetNodeRoots(ByRef list As DataTable) As Collection
Dim rootNodes As New Collection
Dim lastRoot As String = String.Empty
For Each row As DataRow In list.Rows
Dim nodeList() As String = row("LocationName").ToString.Split(CChar("\"))
If lastRoot <> nodeList(0) Then
rootNodes.Add(nodeList(0))
lastRoot = nodeList(0)
End If
Next
Return rootNodes
End Function
Public Function BuildNode(ByVal path As String, ByRef list As DataTable) As TreeNode
' Get the current path where we are upto.
Dim currentPath() As String = path.Split(CChar("\"))
' Create next node.
Dim node As New TreeNode(currentPath(currentPath.GetUpperBound(0)))
' Get a filtered list of all the child nodes for the path passed in.
For Each row As DataRow In list.Select(String.Concat("LocationName LIKE '", String.Concat(path, "\"), "%'"))
' Get the full path of the next node.
Dim nextNode() As String = row("LocationName").ToString.Split(CChar("\"))
' If the next node has any more childs then call the recursive function.
If nextNode.Count > currentPath.Count Then
' Make the next path we pass in just 1 step along from the current path.
ReDim Preserve nextNode(currentPath.Count)
' Call recursive funxtion with the new path, rejoin the path back to its orginal format also.
node.Nodes.Add(BuildNode(String.Join("\", nextNode), list))
End If
Next
Return node
End Function
ASKER
Ok while some essential tweeks to the above code for efficency I stumbled accross the probem,
the problem lies with that line
For Each row As DataRow In list.Select(String.Concat( "LocationN ame LIKE '", path, "\", "%'"))
It works fine except for its returning to much, if the [path] was
LocationName LIKE 'A\1\%'
it works but its returning multipule entries due to there been no GROUP BY or SELECT DISTINCT clauses.
The problem is I only know how to filter the results in DataTable.Select with parameters. Can you do this in the DataTable or do I have to resort to doing it manually in the code itself?
the problem lies with that line
For Each row As DataRow In list.Select(String.Concat(
It works fine except for its returning to much, if the [path] was
LocationName LIKE 'A\1\%'
it works but its returning multipule entries due to there been no GROUP BY or SELECT DISTINCT clauses.
The problem is I only know how to filter the results in DataTable.Select with parameters. Can you do this in the DataTable or do I have to resort to doing it manually in the code itself?
Is it because the select clause returns you all the descendants of a node, whether direct or indirect.
That is when current node is 'A\1'
The select will return you both 'A\1\B', as well as 'A\1\B\2'?
If this is the case your check should be
If nextNode.Count = currentPath.Count + 1 Then
instead of
If nextNode.Count > currentPath.Count Then
That is when current node is 'A\1'
The select will return you both 'A\1\B', as well as 'A\1\B\2'?
If this is the case your check should be
If nextNode.Count = currentPath.Count + 1 Then
instead of
If nextNode.Count > currentPath.Count Then
ASKER
Yes that is whats happening, the select clause is returning multipule 'A\1\' 's due to child nodes.
however
If nextNode.Count = currentPath.Count + 1 Then
does top the endless looping so I can actually see the results but were missing loads of nodes, we only go max 1 depth in now.
however
If nextNode.Count = currentPath.Count + 1 Then
does top the endless looping so I can actually see the results but were missing loads of nodes, we only go max 1 depth in now.
Hmm, how does your code looks like now after some essential tweaks that you mentioned?
ASKER
tbh pretty much the same I just changed how I was building strings, I had some extra steps in there that wasn't needed, which since this would be looping several 1000's times (alot of records in the DataTable) I didnt want to many overheads.
Looking at whats happening or what we want to stop happening. The Select clause is ment to retrive a list of the Childs ware are to build nodes for then pass those Childpaths into the function again but if them childs have subChilds we have duplicate requests. So we need to remove them duplicates since were not intrested in them.
I can only think of 2 ways atm, either Create a function that filders the orginal Array of DataRows() removing all duplicates before we start the for next loop. or do an extra check and verify the previously processed node is not = to the next node were about to process. ?? /shrug :)
Looking at whats happening or what we want to stop happening. The Select clause is ment to retrive a list of the Childs ware are to build nodes for then pass those Childpaths into the function again but if them childs have subChilds we have duplicate requests. So we need to remove them duplicates since were not intrested in them.
I can only think of 2 ways atm, either Create a function that filders the orginal Array of DataRows() removing all duplicates before we start the for next loop. or do an extra check and verify the previously processed node is not = to the next node were about to process. ?? /shrug :)
ASKER
Tried both methods and the overheads are to high takes 5-6Ssecs pn my PC to generate 1011 nodes.
Which is far to long as the final plan would be much larger. The problem is working with the DataTable only. So I'll write it so I can actually query the database for the child nodes and let the database do the filtering.
ty :)
Which is far to long as the final plan would be much larger. The problem is working with the DataTable only. So I'll write it so I can actually query the database for the child nodes and let the database do the filtering.
ty :)
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Ty, I'd orginally tried to do it that way but failed. After having another go at it I realised I was using the TreeView control incorrectly.
Its all working now and the tree is build pretty much instantly even with all the records in.
I've attached the basic code that gets it working.
Thanks. :)
Its all working now and the tree is build pretty much instantly even with all the records in.
I've attached the basic code that gets it working.
Thanks. :)
For Each row As DataRow In data.Rows
Dim nodePath() As String = row("LocationName").ToString.Split(CChar("\"))
Dim currentLocation As TreeNode
Dim nextPath As String = String.Empty
For Each node As String In nodePath
' Keep track of our current path, so we can use it as the Node's Key property.
nextPath = String.Concat(nextPath, node)
If currentLocation Is Nothing AndAlso LocationTreeView.Nodes.Item(nextPath) Is Nothing Then
' Create a new root node.
currentLocation = LocationTreeView.Nodes.Add(node, node)
ElseIf currentLocation IsNot Nothing AndAlso currentLocation.Nodes.Item(nextPath) Is Nothing Then
' Create a new sub node.
currentLocation = currentLocation.Nodes.Add(nextPath, nextPath)
ElseIf currentLocation Is Nothing AndAlso LocationTreeView.Nodes.Item(nextPath) IsNot Nothing Then
' Root node already exists, set our location to root node.
currentLocation = LocationTreeView.Nodes.Item(nextPath)
Else
' Sub node exists, move our current location to subnode.
currentLocation = currentLocation.Nodes.Item(nextPath)
End If
Next
' Reset current location back to root.
currentLocation = Nothing
nextPath = String.Empty
Next
Wow, thanks for sharing with us, and glad that I can help ... thanks for the points too :)
Open in new window